台灣地區登革熱擴散之空間分析

Abstract

近年來地理資訊系統（Geographic Information System, GIS）與流行病學的結合應用，增進我們對於疾病空間特徵的掌握與瞭解。本研究從醫學地理的觀點切入，繪製疾病地圖並探討其空間分布，運用空間自相關的方法進行登革熱聚集與擴散特性之研究，討論登革熱聚集現象，並加入時間的向度，分析登革熱擴散的方式與機制。最後，以人口移動和人口密度作為影響因子，討論此兩項因子與登革熱發生率的關係。 透過空間自相關的檢定，顯示台灣地區的登革熱有明顯的空間聚集現象，且核心區只有一個，以接觸傳染為主要傳播的方式；利用LISA（Local Indicators of Spatial Association）分析不同時間剖面的熱點，再將不同時間剖面的資料串連呈動態的擴散過程，本研究發現在擴散階段，登革熱是以感染式擴散為主要方式；在疫情消退階段，則呈現階層式擴散的型態。有關影響登革熱擴散的因素，在橫斷面分析方法下，研究範圍內的登革熱疫情受氣候影響情形並不顯著；以人口移動作為影響因子進行分析，發現人口移動增強疫情的擴散，在消退的階段影響則不明顯；以人口密度作為影響因子進行分析，發現人口密度也影響登革熱疫情的傳播。

Recent application of Geographic Information System (GIS) in epidemiology research enhances our understanding on the spatial characteristics of diseases. This research generates disease maps and uses spatial autocorrelation to analyze the clustering and diffusion of dengue fevers occurred in Taiwan. The factors being studied include: climate, the movement of population and the density of population. By spatial autocorrelation, this research finds that dengue fever is clustered in Taiwan with one cluster only. Using LISA（Local Indicators of Spatial Association）, this research identifies several hotspots at different time periods. It is also found that the spreading period, contagion diffusion is the major pattern while hierarchical diffusion is the major one when the number of diseases fading out. Under the cross-section analysis, this research finds that climate’s impact on the distribution of dengue fever is not significant within the study area; the movement of population speeds up the diffusion of dengue fever while its influence is less significant when incidence rates fading out. Also, the density of population demonstrates a strong influence on the transmission of dengue fever.